AI news brief: In his latest blog post, Anthropic CEO Dario Amodei echoes calls to ‘slow the pace of frontier AI development’ and lays out three strategies, one of which Anthropic will commit to unilaterally. The backdrop is researcher Jacob Coxon’s resignation from Anthropic this week, citing…
AI News Brief: Fujitsu plans to start exporting AI semiconductors developed using technology related to its Fugaku supercomputer to the US and Asian markets as early as next year. The move is part of Fujitsu’s strategic push to capture growing demand for AI inference CPUs, using its logic semiconductor products to expand its footprint and visibility in the global semiconductor market. The original summary only covers high-level export plans and timing, without disclosing specific chip specs, production capacity, partners, or order details — see the original article for more.
AI news brief: this story covers an Anthropic employee’s resignation and public warning. An Anthropic researcher announced their departure this week and posted a warning on X, accusing the company of racing straight toward self-improving superintelligence while gambling with all of humanity’s lives. Notably, Anthropic’s internal alignment lead didn’t push back or try to smooth things over — instead they co-signed the post, meaning the company’s own internal safety watchdog publicly backed a departing employee’s grim warning.
AI news brief: this is a market research/news translation task, handled directly. Twenty-five Fields Medalists have signed an open letter warning that AI labs racing to claim proofs of famous math problems are threatening the academic norms of the math community. The letter says these AI solutions are often rushed out without time to fully write up the reasoning, clarify new methods and ideas, or properly cite prior related work, sparking serious attribution and plagiarism disputes.
AI news brief: Huawei has rolled out BoGuan, a large language model trained on regional history, in Guangzhou, Guangdong, China, to help generate tourism content and attract visitors. According to reports, the model was trained on historical and cultural data from China’s Shaanxi province, home to well-known attractions like the ancient capital of Xi’an. Its applications target the practical output side of tourism marketing, including helping brainstorm and produce souvenir merchandise and design travel itineraries…
AI news brief: GPT-6 Astra brings a testing upgrade to Cognition’s AI coding assistant Devin, with the core goal of improving Devin’s ability to verify the correctness of its own generated code, so engineers can spend less time on manual code review and ship faster. The update focuses on letting Devin more effectively test software functionality after finishing a dev task and actually prove the code works, instead of just producing code and handing it off for human review.
AI news brief: Perplexity has rolled out the Astra model in production to draft communications, modify software code, and monitor the health of live systems. Compared to the models it used before, Astra needs noticeably less human check-in and approval to do these jobs, meaning the system can run autonomously for longer stretches without constantly stopping for people to review progress or sign off on the next step.
AI news brief: Y Combinator CEO Garry Tan has publicly said he doesn’t want regulators cracking down on Chinese AI labs for ‘distilling’ frontier models, and he’s even arguing the US should build its own distillation pipeline. Distillation means learning how another model works and reasons by querying it extensively, and it’s a common, legal practice in the industry. In a CNBC interview, Tan said…
AI news brief: Mecka AI, a startup that collects and analyzes human motion data to train humanoid and other robotic systems, is close to closing a new funding round led by Sequoia Capital at a valuation of around $500 million, according to two people familiar with the matter. This comes just three months after its last round, led by Framework Ventures with participation from Menlo Ventures, SV Angel, and Kindred Ventures.
AI News Brief: Nscale, a UK-based AI data center startup, has announced that former OpenAI executive Fidji Simo is joining its board, timed right ahead of the company’s planned IPO this fall. Simo will sit on Nscale’s board alongside heavyweight tech leaders like Sheryl Sandberg, Susan Decker, and Nick Clegg. She…